---
type: source
title: "GLP-1 + CBT Reduces Heavy Drinking 41% in RCT — NNT 4.3, Superior to All Approved AUD Medications"
author: "NIH / JAMA Psychiatry (Hendershot et al.)"
url: https://www.nih.gov/news-events/news-releases/adding-weekly-glp-1-cognitive-behavioral-therapy-further-reduces-heavy-drinking
date: 2026-04-01
domain: health
secondary_domains: []
format: research-summary
status: unprocessed
priority: high
tags: [GLP-1, semaglutide, alcohol-use-disorder, behavioral-health, mental-health, clinical-trial, RCT]
intake_tier: research-task
---

## Content

Randomized, double-blind, placebo-controlled clinical trial published in JAMA Psychiatry. 108 patients with AUD + obesity. 26-week duration. Participants received standard cognitive behavioral therapy (CBT) plus either weekly semaglutide or placebo.

**Key results:**
- Semaglutide group: **41.1% reduction in heavy drinking days**
- 13.7% greater improvement than placebo group
- Blood-alcohol biomarkers corroborated self-reported data
- Weight, blood pressure, other clinical measures improved more in semaglutide group
- Gastrointestinal side effects: transient and mild

**Efficacy comparison:**
- **NNT 4.3** for semaglutide (number needed to treat to prevent one heavy drinking day)
- Approved AUD medications (naltrexone, acamprosate): NNT 7 or higher
- Semaglutide NNT is the best in class by a significant margin

**Current landscape:**
- Phase 3 trials evaluating semaglutide for AUD now underway
- A separate low-dose semaglutide trial also showed reductions in laboratory alcohol self-administration and weekly craving (independent replication)
- A pooled meta-analysis of three RCTs showed non-significant association — heterogeneity across study populations may explain

**Safety/complexity:**
- A large community-based cohort study (separate from this RCT) found **195% increased risk of major depressive disorder** among individuals treated with liraglutide or semaglutide
- Researchers emphasize need for comprehensive psychiatric assessment before initiating GLP-1 therapy in at-risk populations
- The depression risk signal is from observational data and may be confounded by indication (obese/metabolically ill patients have higher baseline depression rates)

**NIH quote:** "A new option that is more accessible and more effective could be a gamechanger for closing the treatment gap."

**Full citation:** Hendershot et al., JAMA Psychiatry, 2025. Published February 2025, NIH press release April 2026.

## Agent Notes
**Why this matters:** GLP-1 receptor agonists just demonstrated efficacy for alcohol use disorder at NNT 4.3 — better than every approved AUD medication. This extends GLP-1 therapeutic scope from metabolic health into behavioral/addiction medicine. AUD affects 14M+ US adults and is a major social determinant of health (income loss, family breakdown, mortality). If GLP-1 becomes first-line AUD treatment, it creates a mechanistic bridge between the metabolic health revolution and the behavioral health crisis.

**What surprised me:** The magnitude of the NNT improvement. NNT 4.3 vs. 7+ for approved medications isn't a marginal improvement — it's a category change. The existing medications for AUD (naltrexone, acamprosate) are rarely prescribed despite being effective because of poor NNT. If semaglutide enters the category, prescribing rates could be dramatically higher (it's already prescribed broadly for obesity/diabetes).

**What I expected but didn't find:** A clearer mechanism for the addiction effect. The reward salience hypothesis (GLP-1 reduces the hedonic value of alcohol like it reduces food craving) is the leading theory but not confirmed. This matters for whether the effect extends to other substance use disorders (nicotine, cocaine).

**KB connections:**
- the mental health supply gap is widening not closing — GLP-1 for AUD is a pharmacological bypass of the workforce constraint (no therapist needed for prescribing pathway)
- [[GLP-1 receptor agonists are the largest therapeutic category launch in pharmaceutical history but their chronic use model makes the net cost impact inflationary through 2035]] — AUD indication could expand the market dramatically beyond metabolic disease
- medical care explains only 10-20 percent of health outcomes because behavioral social and genetic factors dominate — AUD is a behavioral/social health driver; pharmacological treatment of AUD via GLP-1 would address a non-clinical determinant through clinical means

**Extraction hints:**
- Strong new claim candidate: "GLP-1 receptor agonists demonstrate NNT 4.3 for alcohol use disorder — superior to all approved AUD medications — extending GLP-1 therapeutic scope from metabolic to behavioral health"
- Note the complication: 195% MDD risk from cohort study must be acknowledged as challenged_by in the claim
- The AUD + obesity comorbidity is the studied population — scope carefully (this is not general population AUD, but obese + AUD, which is ~40% of AUD patients)
- Cross-domain: behavioral health + metabolic intersection

**Context:** First RCT evidence for a GLP-1 agonist in AUD treatment. Phase 3 trials will determine whether this reaches clinical guidelines. The NNT advantage is significant because existing AUD medications are under-prescribed — semaglutide's broad adoption in obesity/diabetes could translate to dramatically higher AUD treatment penetration.

## Curator Notes (structured handoff for extractor)
PRIMARY CONNECTION: GLP-1 receptor agonists are the largest therapeutic category launch in pharmaceutical history... — this extends the therapeutic scope claim
WHY ARCHIVED: First RCT evidence of GLP-1 for AUD; NNT 4.3 vs. 7+ approved medications is a category-level finding, not an incremental update
EXTRACTION HINT: Write as a new claim scoped to "in adults with comorbid AUD and obesity" — do not generalize to all AUD patients. Acknowledge the cohort study MDD risk signal as challenged_by. Flag for Clay (narrative: substance use has major cultural/social dimensions) and Theseus (behavioral AI safety analog: treating behavioral patterns pharmacologically).